Alerts

Alerts are notifications that you can set up to inform you about an event. They can be created via the Create Notification selection on the Notifications page under Alerts or can be defined in the new event definition workflow.

Assigned alerts are displayed on the Notifications page. This page is also bulk friendly and allows you to edit multiple entities. You can see if your alerts are active by clicking on the Test Notifications button under More Actions. You will then see a success or error message under the entity title.

In this section, we explain how to create an alert and how to configure supported alert types.

Hint: Alerts are meant to be extensible through plugins. You can find more options in Graylog Marketplace or even create your own.

Prerequisites

The following prerequisites are required before utilizing alerts with Graylog:

Alerts can be created by selecting Notifications under the Alerts menu or by defining them in the event workflow.

Events and Alerts Page

The filter functionality on the Alerts & Events page in the General perspective allows you to quickly drill down into specific types of alerts by applying targeted filters. This helps narrow the scope of your active alerts, making it easier to focus on relevant data during monitoring or investigation.

Search Panel

At the top of the page, you'll find a search bar designed to help you search through events by using keywords. To further drill down on search results, you can use filters to find specific events or alerts.

Filters

Filters allow you to narrow down events or alerts with precision. Select the + icon dropdown button next to Filters to view a list of available filter types:

Filter Option Purpose
Type Filters for event categories which could either be and event or an alert.
Priority Shows only events with selected severity levels (e.g., Low, Medium, High).
Timestamp Defines a a time window to filter events within a date/time range. Useful for incident review.

Event Definition

View only events or alerts based on predefined system or user-defined event rules.

Aggregation time range Time window used for evaluating the event definition.
ID Search by specific event IDs if known.

When a filter is applied, it becomes visible as a tag next to the Filters dropdown menu. Multiple filters can be selected and combined for precision targeting.

Hint: You can only use one filter for a Boolean attribute. For example, you can either filter for an event or an alert but not both at the same time.

Metrics Graph Widget

This widget plots the count of events over time. It helps quickly identify spikes or trends in event occurrences.

  • Blue Line (Events): Events are color-coded to blue and represent detected events.

  • Orange Line (Alerts): Alerts are color-coded to orange and highlights triggered alerts.

  • Interactive Axis: Hover over graph points for a summary of event counts at specific times. X-axis represents a timeline, and the Y-axis represents count.

  • Export: The chart can be exported using the download icon for reporting purposes.

  • Auto-refresh: Select update behavior by clicking the Not updating dropdown and choosing a timeframe ( e.g. 5 seconds, 10 seconds, 1 minute or 5 minutes) for real-time monitoring.

This widget is instrumental in identifying spikes in malicious behavior or assessing alert volumes over time.

If the message “No events have been found” is displayed, it means no event definition rules have triggered within the selected filters or timeframe. You may need to:

  • Expand the timestamp range.

  • Remove filters.

  • Confirm event generation on the backend is functioning.

Beneath the metrics graph, all your events and alerts are listed in a table, grouped into various columns, beginning with a description of the event/alert, priority, key, type, event definition, event definition type, timestamp and actions. Additional columns can be displayed or removed by clicking on the Columns dropdown button and selecting available column categories.

From the list of all displayed alerts and events, you can either click on an event or alert to expand and view its details or you can click on the Details button in the Actions column.. You can also add notifications to an event by clicking on the More dropdown menu button in the Actions column and select Send Notifications.

Alerts Page Example

You can troubleshoot failed inputs by filtering. For instance, an event like INPUT_FAILING: An input has failed to start is a critical system warning which surfaces in the alerts and events page. Filtering by Type = Event, Event Definition = System notificationevents and Priority = Low/High quickly isolates these kinds of events for response or escalation.

Create an Alert from the Alerts Menu

  1. Navigate to the Alerts menu and select Notifications.

  2. Select the Create notification button.

  3. Complete the following fields: 

    1. Title: Create a unique title for your alert.

    2. Description (optional): You may add additional details about your alert in this field if desired.

    3. Notification Type: Select the alert type from the drop-down menu.

  4. After you select an alert type, you are presented with additional fields based on the type selected. The alert types and key fields to configure for each are detailed in Alert Types.

  5. You may also choose to test your alert at this time by selecting Execute Test Notification.

  6. Click Create notification.

Create an Alert in the New Event Definition Workflow

You may also choose to create an alert while you are in the process of defining a new event.

  1. In the New Event Definition menu, there will be a selection for Notifications in the menu bar.

  2. Under Add Notification, select Create New Notification from the drop-down menu.

  3. From the menu that populates, complete the following fields: 

    1. Title: Create a unique title for your alert.

    2. Description (optional): You may add additional details about your alert in this field if desired.

    3. Notification Type: Select the alert type from the drop-down menu.

  4. After you select an alert type, you are presented with additional fields based on the type selected. The alert types and key fields to configure for each are detailed in Alert Types.

  5. You may also choose to test your alert at this time by selecting Execute Test Notification.

  6. Click Create notification.

Metadata Available to Alerts

When creating alerts you can utilize metadata from the event definition, the event itself, and the event's backlog messages (if it is configured to retain a backlog). This metadata can be used when formatting email, Slack, and Microsoft Teams alerts or when providing arguments to a script alert.

For example, if you wish to include more information in your Slack alerts, you may add new fields to the Custom Message section. You may also remove any fields that you do not wish to see by deleting them from this section.

Or you could add arguments to a script alert to include more information in your alerts.

Fields that are available for each entity type are detailed below.

Event Definition Metadata

Field Type Description
event_definition_id

String

The database ID of the event definition
event_definition_type

String

The internal name of the event definition type (aggregation-v1 or correlation-v1)
event_definition_title

String

The title set in the UI
event_definition_description

String

The description set in the UI
job_definition_id

String

The internal job definition ID associated with a scheduled event definition
job_trigger_id

String

The internal ID associated with the current execution of the job.

Event Metadata

Field Type Description
event

 

The event as it is stored in Graylog

id

String

The message ID of the stored event

event_definition_id

String

The database ID of the event definition

event_definition_type 

String

The internal name of the event definition type (aggregation-v1 or correlation-v1)

origin_context

String

URN of the message or event creating this event (eitherevent or message) (can be empty)

timestamp

DateTime

The timestamp can be set to the underlying event or message (see origin_context above)

timestamp_processing

DateTime

The timestamp for when the event was created by Graylog

timerange_start

DateTime

The start of the window of data Graylog used to create this event (can be empty)

timerange_end

DateTime

The end of the window of data Graylog used to create this event (can be empty)

streams

String

The list of stream IDs the event is stored in

source_streams 

String

The list of stream IDs the event pulled data from

alert

Boolean

Whether this event is considered to be an alert; always true for event definitions that have alerts

message 

String

A human-friendly message describing this event

source

String

The host name of the Graylog server that created this event

key_tuple

String

The list of values making up the event’s key

key

String

The event’s key as a single string

priority

Long

The event’s priority value

fields

Map

The custom fields attached to the event

Backlog Metadata

Field Type Description
backlog

 

The list of messages or events which lead to the alert being generated

id 

String

The message ID

index

String

The name of the index the message is stored in; use together with id to uniquely identify a message in Graylog

source

String

The source field of the message

message

String

The message field of the message

timestamp

DateTime

Thetimestamp field of the message

stream_ids

String

The stream IDs of the message

fields

Map

The remaining fields of the message (can be iterated)

Delete Queued Alerts

If processing stops and event updates begin to pile up in the queue, then you might have unknowingly fired too many alerts. To avoid an influx of alerts, make sure to set an alert grace period for event definitions. The grace period enforces a rate limit on how many alerts are triggered for identical events. This effectively prevents queued event alerts. Without a grace period in place, too many event triggers can cause a backlog of alerts.

If you are faced with queued event alerts, there are two ways of clearing the alert queue.

Clear Alert Queue Manually

Clear the alert queue manually through the interface:

  1. Navigate to the Events Definition menu by selecting Alerts > Event Definitions.

  2. From the list of definitions available, click on the Information icon under Scheduling.

  3. The event definition menu will expand. Here, you will see the number of queued alerts. If there are a lot of queued alerts, this typically suggests an abnormality. On the Queued alerts line, click on clear to clear queued alerts for the selected event definition.

Disable an Event

You can also clear the alert queue by disabling an event.

  1. As in the previous example, navigate to the Events Definitions tab.

  2. Next to your event definition, click the More drop-down button, and select Disable from the menu option for the event you wish to disable.

  3. Upon selecting the Disable option, a pop-up dialog screen appears, prompting you to confirm the selection.

  4. When disabled an alert is displayed confirming that the selected event definition has been disabled.